Constructing models for Crohn's disease diagnosis and prediction of infliximab non-response based on angiogenesis-related genes

Abstract

Background: Angiogenesis response plays a crucial role in the occurrence and development of Crohn's disease (CD) and may involve the mechanism of infliximab non-response. However, the role of angiogenesis-related genes in Crohn's disease has not been comprehensively studied. This study aimed to explore the expression profiles of angiogenesis-related genes in CD patients and construct models for disease diagnosis and prediction of infliximab non-response.

Methods: CD-related microarray datasets were collected from the GEO database. Unsupervised consensus clustering analysis was performed based on differentially expressed angiogenesis-related genes to divide CD samples into two distinct clusters. Weighted gene co-expression network analysis (WGCNA) was conducted on the clusters to identify angiogenesis-related module. Based on the differentially expressed genes in the module, machine learning algorithms were employed to further identify hub genes and construct a disease diagnostic model. Subsequently, treatment outcome-related genes were extracted from these hub genes, and a predictive model for infliximab non-response in CD patients was ultimately built.

Results: Based on angiogenesis-related genes, we identified two distinct CD clusters (C1 and C2). Compared to C1, the metabolic pathways in C2 were significantly upregulated, and there was a higher abundance of cell clusters such as M1 macrophages and plasma cells. Additionally, C2 showed a poorer response to infliximab. Furthermore, a predictive model for infliximab non-response in CD patients was constructed based on the hub genes, and it was successfully validated using an external dataset.

Conclusion: Comprehensive analysis of angiogenesis-related genes revealed different clusters of CD, which exhibited differential response rates to infliximab. The construction of models provides a reference for disease diagnosis and drug selection, aiding in clinical decision-making.

Keywords: Crohn’s disease; angiogenesis; bioinformatics analysis; infliximab; machine learning; prediction model.